This past week, the AI world is witnessing a significant milestone with the release of Meta AI's new Llama 2, an advanced open-source language model available for both research and commercial applications.

📢 AI Spotlight: Meta AI's Llama 2 hits the stage - A more powerful language model that broadens the horizon for commercial AI use. 

🔍 AI in Action: GitLab Duo - Accelerating DevSecOps Through Generative AI

Here’s what we have today:

📢 AI Radar

1. Meta AI's new Llama 2, an upgraded open-source language model, shows improved reasoning and coding skills. Trained on 2 trillion tokens for longer context, it's available free for both research and commercial applications. (link)

2. Google Bard was updated with some major upgrades. It’s now accessible in over 40 languages. The latest update allows you to customize responses, add images to prompts, and it's now available in more places, including Brazil and Europe. The experience is designed to boost creativity and productivity, enabling more dynamic interactions with AI. (link)

3. Apple is developing its own AI-powered chatbot, internally known as "Apple GPT", built on its large language model framework, "Ajax". Although the company's plans for public release remain unclear, the technology could potentially rival similar tools from OpenAI, Google, Microsoft, and Meta. This marks Apple's significant exploration into the generative AI space, further enhancing its AI portfolio alongside Siri. (link)

4. LangChain. the popular language model development framework, has unveiled LangSmith, a platform designed to simplify the process of transitioning large language model applications from prototype to production. The platform focuses on facilitating debugging, testing, evaluating, and monitoring of LLM applications, and is currently in closed beta. (link)

5. OpenAI is introducing custom instructions for its large language model, ChatGPT, allowing users to add preferences or requirements to the AI's responses. The new feature will first be available in beta with the Plus plan and expand to all users in the coming weeks. (link)

🔍 AI in Action - GitLab Duo: Accelerating DevSecOps Through Generative AI

GitLab, a renowned DevOps platform, serves as a unified workspace for software development, assisting companies around the globe. With the ability to manage and streamline project planning, source code management, CI/CD, monitoring, and security all in one place, GitLab significantly accelerates software delivery while improving operational efficiency.

AI at GitLab

GitLab has been integrating AI to augment its functionalities and enhance user experience. Generative AI has proven to be a game changer in the realm of software development. A study by Peng et al. in 2023 found that developers using AI-powered tools completed their tasks 55.8% faster than those who didn’t. Embracing this efficiency, GitLab introduced GitLab Duo, a suite of AI capabilities designed to boost productivity for DevSecOps teams throughout the software development lifecycle.

Focusing on AI-Driven Efficiency and Secure Development

GitLab's AI deployment emphasizes several key areas, including optimizing developer productivity, ensuring secure software development, and enhancing the code review process.

Boosting Developer Productivity: GitLab Duo utilizes the power of generative AI to offer features like Code Suggestions and Code Explanations. These AI-powered tools provide real-time, accurate coding suggestions and explain code blocks in natural language, greatly boosting developer productivity. GitLab Duo also leverages AI to generate test files for code changes, reducing the time developers spend writing tests, thereby enhancing efficiency.

Ensuring Secure Software Development: Alongside productivity, GitLab Duo also prioritizes secure software development. The suite offers detailed explanations for identified vulnerabilities, encompassing their context, potential exploitation methods, and possible fixes. This quick vulnerability remediation capability strengthens software security.

Enhancing Code Review Processes

In a 2019 survey, 55% of respondents were indifferent to or dissatisfied with their team’s code review process. Addressing this issue, GitLab employs AI to revolutionize traditional code review processes. GitLab Duo offers AI-powered capabilities like Merge Request Change Summary and Issue Comment Summary, providing insightful summaries of merge requests and issue comments. The suite can also suggest potential reviewers for a merge request, ensuring higher-quality feedback and accelerating review processes.

Looking Forward

GitLab continues to innovate with AI. The company is piloting new features such as GitLab Duo Chat, which uses AI to answer product-specific questions about GitLab, reducing time spent on reviewing documentation. Another experiment is the "fill in merge request description" feature, which leverages AI to populate merge request descriptions based on proposed changes. With ongoing advancements in its AI-assisted workflows, GitLab is paving the way for teams to deliver secure, high-quality software faster and with increased efficiency.

That's all for today! If you’re looking to accelerate your business growth with AI:

Don't forget to keep an eye on this newsletter for more cool stories about how companies like GitLab use AI.